Publication number | US8009382 B2 |

Publication type | Grant |

Application number | US 12/565,382 |

Publication date | Aug 30, 2011 |

Filing date | Sep 23, 2009 |

Priority date | Apr 27, 2007 |

Fee status | Paid |

Also published as | US20100014184, WO2008139505A1 |

Publication number | 12565382, 565382, US 8009382 B2, US 8009382B2, US-B2-8009382, US8009382 B2, US8009382B2 |

Inventors | Kazuhiko Takaishi |

Original Assignee | Toshiba Storage Device Corporation |

Export Citation | BiBTeX, EndNote, RefMan |

Patent Citations (18), Referenced by (1), Classifications (9), Legal Events (4) | |

External Links: USPTO, USPTO Assignment, Espacenet | |

US 8009382 B2

Abstract

According to one embodiment, a position control apparatus includes: a digital control module that performs position control including disturbance adaptive control with a predetermined sampling period using a control constant, in accordance with a position error; and a table that stores the control constant corresponding to one of values sin(ωT), cos(ωT), and sin(ωT/2), wherein the digital control module determines a control value of the actuator in accordance with the position error, calculates a value of one of estimated errors of sin(ωT), cos(ωT), and sin(ωT/2) according to an adaptive law from a signal based on the position error, multiplies the calculated estimated error by an adaptive gain in accordance with a value of the angular frequency ω, updates one of the values, reads the corresponding control constant from the table based on the updated value, and updates the control constant.

Claims(20)

1. A position control apparatus that performs position control on an object using an actuator in accordance with a position of the object, comprising:

a digital control module configured to perform position control including disturbance adaptive control with a predetermined sampling period using a control constant, in accordance with a position error between a target position and a current position of the object; and

a table configured to store the control constant corresponding to one of values sin(ωT), cos(ωT), and sin(ωT/2), with ω representing an angular frequency of a disturbance, T representing the sampling period,

wherein the digital control module determines a control value of the actuator in accordance with the position error, calculates a value of one of estimated errors of sin(ωT), cos(ωT), and sin(ωT/2) according to an adaptive law from a signal based on the position error, multiplies the calculated estimated error by an adaptive gain in accordance with a value of the angular frequency ω, updates one of the values sin(ωT), cos(ωT), and sin(ωT/2), reads the corresponding control constant from the table based on the updated value, and updates the control constant, to perform the position control including the disturbance adaptive control.

2. The position control apparatus of claim 1 , wherein the digital control module performs the position control including the disturbance adaptive control by observer control.

3. The position control apparatus of claim 2 , wherein the digital control module calculates an estimated error from a state variable of the observer, an estimated gain, and an estimated position error, multiplies the estimated error by the adaptive gain, adds the estimated error to one of sin(ωT), cos(ωT), and sin(ωT/2), to update the one of sin(ωT), cos(ωT), and sin(ωT/2).

4. The position control apparatus of claim 2 , wherein the digital control module calculates an estimated error from a state variable of a current sample and a state variable of a next sample of the observer, multiplies the estimated error by the adaptive gain, adds the estimated error to one of sin(ωT), cos(ωT), and sin(ωT/2), to update the one of sin(ωT), cos(ωT), and sin(ωT/2).

5. The position control apparatus of claim 2 , wherein

the table stores a state estimating gain of the observer corresponding to the value of one of estimated errors of sin(ωT) cos(ωT), and sin(ωT/2), and

the digital control module updates the state estimating gain of the observer, as the control constant, by a state estimating gain read from the table.

6. The position control apparatus of claim 1 , wherein the digital control module clips the one of the calculated values sin(ωT), cos(ωT), and sin(ωT/2), so as to restrict the calculated value of the one of sin(ωT), cos(ωT), and sin(ωT/2) within a predetermined range.

7. The position control apparatus of claim 1 , wherein

the digital control module includes:

a control module configured to calculate the control value of the actuator in accordance with the position error; and

a disturbance adaptive control module configured to calculate the one of values sin(ωT), cos(ωT), and sin(ωT/2) according to the adaptive law, multiplies the calculated value by the adaptive gain to update the one of sin(ωT), cos(ωT), and sin(ωT/2), and calculates a disturbance adaptive control value using the control constant that is read from the table in accordance with the updated value, and

the actuator is driven with the control value of the control module and the disturbance adaptive control value.

8. The position control apparatus of claim 1 , wherein the digital control module multiplies the value of sin(ωT) by an adaptive gain that is proportional to cos(ωT).

9. The position control apparatus of claim 1 , wherein the digital control module multiplies the value of cos(ωT) by an adaptive gain that is proportional to sin(ωT).

10. The position control apparatus of claim 1 , wherein the digital control module multiplies the value of sin(ωT/2) by an adaptive gain that is proportional to cos(ωT/2).

11. A disk apparatus comprising:

a head configured to read information from a disk;

an actuator configured to move the head in a transverse direction of tracks of the disk;

a digital control module configured to perform position control including disturbance adaptive control with a predetermined sampling period using a control constant, in accordance with a position error between a target position and a current position of the object; and

a table configured to store the control constant corresponding to one of values sin(ωT), cos(ωT), and sin(ωT/2), with ω representing an angular frequency of a disturbance, T representing the sampling period,

wherein the digital control module determines a control value of the actuator in accordance with the position error, calculates a value of one of estimated errors of sin(ωT), cos(ωT) and sin(ωT/2) according to an adaptive law from a signal based on the position error, multiplies the calculated estimated error by an adaptive gain in accordance with a value of the angular frequency ω, updates one of the values sin(ωT), cos(ωT), and sin(ωT/2), reads the corresponding control constant from the table based on the updated value, and updates the control constant, to perform the position control including the disturbance adaptive control.

12. The disk apparatus of claim 11 , wherein the digital control module performs the position control including the disturbance adaptive control by observer control.

13. The disk apparatus of claim 12 , wherein the digital control module calculates an estimated error from a state variable of the observer, an estimated gain, and an estimated position error, multiplies the estimated error by the adaptive gain, adds the estimated error to one of sin(ωT), cos(ωT), and sin(ωT/2), to update the one of sin(ωT), cos(ωT), and sin(ωT/2).

14. The disk apparatus of claim 12 , wherein the digital control module calculates an estimated error from a state variable of a current sample and a state variable of a next sample of the observer, multiplies the estimated error by the adaptive gain, adds the estimated error to one of sin(ωT), cos(ωT), and sin(ωT/2), to update the one of sin(ωT), cos(ωT), and sin(ωT/2).

15. The disk apparatus of claim 12 , wherein

the table stores a state estimating gain of the observer corresponding to the value of one of estimated errors of sin(ωT) cos(ωT), and sin(ωT/2), and

the digital control module updates the state estimating gain of the observer, as the control constant, by a state estimating gain read from the table.

16. The disk apparatus of claim 11 , wherein the digital control module clips the one of the calculated values sin(ωT), cos(ωT) and sin(ωT/2), so as to restrict the calculated value of the one of sin(ωT), cos(ωT), and sin(ωT/2) within a predetermined range.

17. The disk apparatus of claim 11 , wherein

the digital control module includes:

a control module configured to calculate the control value of the actuator in accordance with the position error; and

a disturbance adaptive control module configured to calculate the one of values sin(ωT), cos(ωT), and sin(ωT/2) according to the adaptive law, multiplies the calculated value by the adaptive gain to update the one of sin(ωT), cos(ωT), and sin(ωT/2), and calculates a disturbance adaptive control value using the control constant that is read from the table in accordance with the updated value, and

the actuator is driven with the control value of the control module and the disturbance adaptive control value.

18. The position control apparatus of claim 11 , wherein the digital control module multiplies the value of sin(ωT) by an adaptive gain that is proportional to cos(ωT).

19. The position control apparatus of claim 11 , wherein the digital control module multiplies the value of cos(ωT) by an adaptive gain that is proportional to sin(ωT).

20. The position control apparatus of claim 11 , wherein the digital control module multiplies the value of sin(ωT/2) by an adaptive gain that is proportional to cos(ωT/2).

Description

This application is a continuation of PCT international application Ser. No. PCT/JP2007/000474 filed on Apr. 27, 2007 which designates the United States, incorporated herein by reference.

1. Field

The present invention relates to a position control apparatus and a disk apparatus for preventing displacement of an object due to a disturbance, and more particularly, to a position control apparatus and a disk apparatus for preventing displacement due to an unknown disturbance frequency.

2. Description of the Related Art

In recent years, devices for performing position control on objects with high precision are widely used. For example, in a disk apparatus that performs recording and reproduction of data on a rotating storage medium (such as a disk medium) with a head, the head is positioned onto a desired track of the storage medium, and data is read from or written on the track. In such a disk apparatus as a magnetic disk apparatus or an optical disk apparatus, it is very important to accurately position a head onto a target track, so as to achieve higher recording density.

One of the factors that hinder the positioning is the decentering caused when the centers of concentric servo signals on a disk differ from the rotational center of the motor. To correct the decentering, a control method involving an observer has been suggested (see Japanese Patent (KOKAI) No. 3,460,795, for example).

The decentering leads to sinusoidal position fluctuations that are in synchronization with an integral multiple of the rotation frequency of the disk. By the above mentioned observer control method, such sinusoidal position fluctuations can be restrained, and accurate positioning on each target track can be performed. However, to perform such a decentering correction, the frequency to be corrected should be known in advance. For example, the frequency to be corrected should be an integral multiple of the rotation frequency, or should be equal to or twice as high as the rotation frequency.

The second factor that hinders the positioning is the vibrations transmitted from outside the disk apparatus. The vibrations may have various waveforms, but in the following, sinusoidal vibrations are explained. Besides, the above decentering correction control may be applied for the case with frequencies that are not integral multiples of the rotation frequency.

In the above conventional structure, the frequency of a disturbance is of a known value. However, external vibrations transmitted from outside are unknown at the stage of designing the control system, and the frequency cannot be known in advance. Therefore, it is necessary to detect unknown frequencies. As long as the frequencies can be detected, position fluctuations due to external vibrations can be prevented (by adaptively following the external vibrations) according to the control method as disclosed in Japanese Patent (KOKAI) No. 3,460,795.

**120** calculates the position error e between a target position r and the current position y, and a control module (Cn) **121** calculates a control amount Un. With this control amount Un, a plant **122** (such as a VCM of a disk apparatus) is driven. In a disk apparatus, servo signals from a magnetic head are demodulated, and the current position y of the plant **122** is calculated and is fed back to the arithmetic module **120**.

An estimator **124** estimates the angular frequency ω of an external vibration, with the use of the position error e and an internal variable of a disturbance suppression compensator **123** (Cd). A compensation table **125** stores constants of the compensator **123** (Cd) for suppressing external oscillations at each frequency ω. The disturbance suppression compensator **123** (Cd) corrects the internal constant in accordance with a constant that is read from the compensation table **125** in accordance with the angular frequency ω of the estimator **124**. In this manner, a disturbance suppression control amount Ud is calculated from the position error e. An addition module **126** adds the control amount Un and the disturbance suppression control amount Ud, and outputs the addition result to the plant **122**.

As described above, according to a conventional method, the angular frequency (the disturbance frequency) ω is estimated, and the internal constant of the compensator Cd is corrected in accordance with the value of the angular frequency ω. In this manner, an optimum operation of the compensator Cd is maintained over a wide frequency range (see Japanese Patent Application Publication (KOKAI) No. 2007-004946, for example).

To adapt to unknown disturbance frequencies, the estimated angular frequency ω is corrected whenever necessary in the conventional estimation of the angular frequency ω. Therefore, the adaptive law is obtained through an analog control system in the following manner. The rotation vector of the compensator that compensates periodic disturbances is expressed by the following series of equations (1):

*z*1=*A*·cos(ω*t*),

*z*2=*A*·sin(ω*t*). (1)

In the equations (1), z1 and z2 represent the X-axis component and the Y-axis component of the rotation vector.

The following equations (2) are obtained by temporally differentiating z1 and z2 of the equations (1) with respect to time. Here, ω is function of time.

The angle of the rotation vector is the tangent of the sin component z1 and the cos component z2, and the angular frequency ω is the differential value of the angle. Accordingly, the angular frequency ω of the rotation vector is determined by the following equation (3):

Here, the compensator uses the following equation (4). In the equation (4), ω is assumed to be a constant.

In the equation (4), z1 represents the sine component of the disturbance, z2 represents the cosine component of the disturbance, L4 and L5 represent input gains, e represents the position error, and s represents the Laplacian.

The state equation of the analog compensator of the equation (4) is expanded, and the differentials of z1 and z2 are determined. The following equation (5) is obtained by substituting the differentials of z1 and z2 into the equation (3).

If the estimated angular frequency of an unknown disturbance is correct, the compensator **123** (Cd) can properly suppress the disturbance. As a result, the position error e or the estimated position error of the observer becomes zero. In other words, when the angular frequency ω to be processed by the compensator **23** (Cd) is the same as the estimated angular frequency ω of a disturbance, ω′ of the equation (5) or the term of the position error e of the right side should be zero. Accordingly, the equation (5) is expressed by a time differential form of the angular frequency, to obtain the adaptive law (an integral compensation law) expressed by the following equation (6):

The value of ω is corrected whenever necessary, with the use of this equation. The equation (6) is transformed into an integral form expressed by a digital control equation, and the following equation (7) is obtained:

Here, K represents the adaptive gain. The equation (7) uses the adaptive law in an addition formula. This can be expressed by a multiplication formula. With the use of this adaptive law, a linear relationship can be obtained between a disturbance oscillation frequency and the estimated frequency ω, as illustrated in

In recent years, the aforementioned control system is applied to a digital control system so as to execute the above control system in data processing of a processor. In the above conventional technique, an analog adaptive law is obtained based on analog formulas, and the analog adaptive law is converted into a digital adaptive law. As a result, when the equation (7) is used in a digital control operation, a disturbance cannot be properly suppressed (or position following cannot be performed with respect to the disturbance) through digital disturbance adaptive control. This is because the estimated angular frequency ω is based on analog formulas, though it is directly determined.

A general architecture that implements the various features of the invention will now be described with reference to the drawings. The drawings and the associated descriptions are provided to illustrate embodiments of the invention and not to limit the scope of the invention.

Various embodiments according to the invention will be described hereinafter with reference to the accompanying drawings. In general, according to one embodiment of the invention, a position control apparatus that performs position control on an object using an actuator in accordance with a position of the object, includes: a digital control module configured to perform position control including disturbance adaptive control with a predetermined sampling period using a control constant, in accordance with a position error between a target position and a current position of the object; and a table configured to store the control constant corresponding to one of values sin(ωT), cos(ωT), and sin(ωT/2), with ω representing an angular frequency of a disturbance, T representing the sampling period. The digital control module determines a control value of the actuator in accordance with the position error, calculates a value of one of estimated errors of sin(ωT), cos(ωT), and sin(ωT/2) according to an adaptive law from a signal based on the position error, multiplies the calculated estimated error by an adaptive gain in accordance with a value of the angular frequency ω, updates one of the values sin(ωT), cos(ωT), and sin(ωT/2), reads the corresponding control constant from the table based on the updated value, and updates the control constant, to perform the position control including the disturbance adaptive control.

According to another embodiment of the invention, a disk apparatus includes: a head configured to read information from a disk; an actuator configured to move the head in a transverse direction of tracks of the disk; a digital control module configured to perform position control including disturbance adaptive control with a predetermined sampling period using a control constant, in accordance with a position error between a target position and a current position of the object; and a table configured to store the control constant corresponding to one of values sin(ωT), cos(ωT), and sin(ωT/2), with ω representing an angular frequency of a disturbance, T representing the sampling period. The digital control module determines a control value of the actuator in accordance with the position error, calculates a value of one of estimated errors of sin(ωT), cos(ωT), and sin(ωT/2) according to an adaptive law from a signal based on the position error, multiplies the calculated estimated error by an adaptive gain in accordance with a value of the angular frequency ω, updates one of the values sin(ωT), cos(ωT), and sin(ωT/2), reads the corresponding control constant from the table based on the updated value, and updates the control constant, to perform the position control including the disturbance adaptive control.

The following is a description of a disk apparatus, disturbance frequency estimation of a position control system, adaptive control, a first embodiment of a position control system, a second embodiment of a position control system, a third embodiment of a position control system, a fourth embodiment of a position control system, a fifth embodiment of a position control system, a sixth embodiment of a position control system, and other embodiments, with reference to the accompanying drawings.

**4** that are magnetic storage medium are provided around a rotating shaft **2** of a spindle motor **5**. The spindle motor **5** rotates the magnetic disks **4**. An actuator (VCM) **1** has magnetic heads **3** at the top ends of its arms, and moves the magnetic heads **3** in the radial direction of the magnetic disks **4**.

The actuator **1** is formed with a voice coil motor (VCM) that revolves around a rotating shaft. In the drawing, the two magnetic disks **4** are mounted on the magnetic disk apparatus, and the four magnetic heads **3** are simultaneously driven by the same actuator **1**.

Each of the magnetic heads **3** includes a read element and a write element. Each of the magnetic heads **3** is formed by stacking the read element including magnetoresistive (MR) elements on a slider and the write element including write coils on the read element.

A position detect circuit **7** converts positional signals (analog signals) read by the magnetic heads **3** into digital signals. A read/write (R/W) circuit **10** controls reading and writing of the magnetic heads **3**. A spindle motor (SPM) drive circuit **8** drives the spindle motor **5**. A voice coil motor (VCM) drive circuit **6** supplies a driving current to the voice coil motor (VCM) **1**, to drive the VCM **1**.

A microcontroller (MCU) **14** detects (demodulates) the current position from the digital positional signal supplied from the position detect circuit **7**, and calculates a VCM drive command value in accordance with the position error of the detected current position and a target position. In other words, the MCU **14** performs position demodulation and servo control. A read only memory (ROM) **13** stores the control program for the MCU **14** and the like. A random access memory (RAM) **12** stores data to be used for the processing by the MCU **14** and the like.

A hard disk controller (HDC) **11** determines a position in a cycle, based on the sector number of servo signals, and records or reproduces data. A random access memory (RAM) **15** for buffering temporarily stores read data and write data. The HDC **11** communicates with a host via an interface (IF) such as a USB, ATA, or SCSI. A bus **9** connects those components.

As illustrated in **16** are circumferentially arranged at equal intervals on each of the tracks from the outer circumference to the inner circumference of each magnetic disk **4**. Each of the tracks is formed with sectors, and the solid lines in **16**. As illustrated in

With the use of the track number “GRAY CODE” and the offset information PosA, PosB, PosC, and PosD, the position of the magnetic head in the radial direction is detected. Based on the index signal “INDEX”, the position of the magnetic head in the circumferential direction is grasped.

For example, the sector number at which an index signal is detected is set at “0”, and the sector number is incremented every time a servo signal is detected. In this manner, the sector number of each sector of each track is obtained. The sector numbers of servo signals are used as the references when data recording or reproduction is performed. An index signal is provided for each one rotation. Instead of the index signals, the sector number may be used.

**14** of **14** checks the position of the actuator through the position detect circuit **7** of **14** performs a servo calculation, and supplies an appropriate current to the VCM **1**. **3** is moved from a track position to a target track position. **1**, the velocity of the actuator (the head), and the position of the actuator (the head).

The seek control transits to coarse control, to settling control, and to following control, so that the head can be moved to the target position. The coarse control is basically velocity control, and the settling control and the following control are basically position control. In either case, it is necessary to detect the current position of the head.

To check the position, the servo signals are recorded on the magnetic disk in advance, as illustrated in **7** converts the servo signals into digital values.

**14**. More specifically, an arithmetic module **20** calculates the position error e between a target position r and a current position y, and inputs the position error e to a control module (Cn) **21**. Based on the position error e, the control module **21** performs a control operation to calculate a control amount Un and drive the VCM **1** and **3** that are a plant (P) **22**. The position of the plant demodulates the servo signals from the magnetic heads **3**, calculates the current position y, and feeds the current position y back to the arithmetic module **20**.

An estimator **24** estimates sin(ωT) with the angular frequency ω of external oscillation, using the position error e and the internal variable of a disturbance suppression compensator (Cd) **23**. A control variable table **25** stores the constants of the compensator (Cd) **23** for suppressing the external oscillation, in association with sin(ωT) with the angular frequency ω.

The disturbance suppression compensator (Cd) **23** corrects the internal constant according to the constant read from the compensation table **25**, based on the sin(ωT) with the angular frequency ω of the estimator **24**. The disturbance suppression compensator (Cd) **23** calculates a disturbance suppression control amount Ud from the position error e. An addition module **26** adds the control amount Un to the disturbance suppression control amount Ud, and outputs the addition result to the plant **22**. The disturbance suppression control amount is the amount to be used for controlling disturbance in such a manner that the positional relationship between the disk and the head is not changed. Since the head is normally displaced by disturbance, adaptive control is performed on the driving value of the VCM that drives the head.

In the embodiments, the estimator **24** estimates sin(ωT). The reason is described in the following. In the following description, the control module **21**, the compensator **23**, and the estimator **24** constitute a control system formed with an observer. First, the adaptive law of a predictive observer is described.

The compensator **23** in the predictive observer is expressed by the following equation (8). In the equation (8), z1 represents the sine component of the disturbance, z2 represents the cosine component of the disturbance, L4 and L5 represent input gains, and e represents the position error.

If the disturbance model corresponds to a sinusoidal wave, the matrix A of the equation (8) is expressed by the following equation (9).

When the equation (8) is expanded, the equation (10) is obtained.

Since the time differential of the equation (2) in analog is a subtraction in digital, the equation (11) is obtained by subtracting Z2[k] from Z2[k+1] of the equation (10), and multiplying the resultant by Z1[k].

*z*1*[k]*(*z*2*[k+*1*]−z*2*[k]*)=*z*1*[k]*(*a*21·*z*1*[k]+*(*a*22−1)*z*2*[k]+L*5·*e[k]*) (11)

Likewise, the equation (12) is obtained by subtracting Z1[k] from Z1[k+1] of the equation (10), and multiplying the resultant by Z2[k].

*z*2*[k]*(*z*1*[k+*1*]−z*1*[k]*)=*z*2*[k]*(*a*12*·z*2*[k]+*(*a*11−1)*z*1*[k]+L*4·*e[k]*) (12)

Since the analog differentiation is a digital subtraction as in the equation (3), the equation (13) is obtained by determining a difference between the equation (11) and the equation (12).

*z*1*[k]*(*z*2*[k+*1*]−z*2*[k]*)−*z*2*[k]*(*z*1*[k+*1*]−z*1*[k]*)=*z*1*[k]*(*a*21·*z*1*[k]+*(*a*22−1)*z*2*[k]+L*5·*e[k]*)−*z*2*[k]*(*a*12·*z*2*[k]+*(*a*11−1)*z*1*[k]+L*4·e[k])=*a*21·*z*1*[k]* ^{2}+(*a*22−*a*11)*z*1*[k]·* *z*2*[k]−a*12·*z*2*[k]* ^{2}+(*L*5·*z*1*[k]−L*4·*z*2*[k]*)*e[k]* (13)

When the disturbance model corresponds to a sinusoidal wave, the following relationships are satisfied as in the equation (9): a11=a22(=cos(ωT)) and a12=−a22(=sin(ωT)). Accordingly, the equation (13) can be transformed into the following equation (14).

*z*1*[k]*(*z*2*[k+*1*]−z*2*[k]*)−*z*2*[k]*(*z*1*[k+*1*]−z*1*[k]*)=sin(ω*T*)(*z*1*[k]* ^{2} *+z*2*[k]* ^{2})+(*L*5·*z*1*[k]−L*4·*z*2*[k]*)*e[k]* (14)

The equation (15) is obtained by dividing the equation (14) by the sum of the square of z1 and the square of z2.

Compared with the equation (5), sin(ωT) is estimated as a disturbance in the equation (15), while ω is estimated as a disturbance in the equation (5). More specifically, in a digital control operation, it is necessary to estimate sin(ωT), and, according to the adaptive law, sin(ωT) is estimated so that the second term of the right side of the equation (15) becomes “0”.

Next, an adaptive law of a current observer is described. A current observer is more widely used than the predictive observer in digital control.

The compensator **23** in the current observer is expressed by the following equations (16). In the equations (16) z1 represents the sine component of the disturbance, z2 represents the cosine component of the disturbance, L4 and L5 represent input gains, and e represents the position error.

The equation (17) is obtained by combining the two equations (16).

If the disturbance model corresponds to the sinusoidal wave, the matrix A of the equation (17) is expressed by the following equation (18). To simplify the equations, cos(ωT) and sin(ωT) will be hereinafter represented by “c” and “s”.

The inverse matrix of the matrix A of the equation (18) is expressed by the following equation (19):

The equation (20) is obtained by dividing the equation (17) by the matrix A of the equation (18), and substituting the inverse matrix of the matrix A of the equation (19) into the resultant.

The equation (21) is obtained by determining z1 and z2 of the k sample from the equation (20).

The equation (22) is obtained by multiplying z1[k] of the equation (21) by z2[k+1], and multiplying z2[k] of the equation (21) by z1[k+1]. Here, the left-side negative sign “−” is given to z2[k].

The equation (23) is obtained by summing up the upper and lower sides of the equation (22).

*z*1*[k]z*2*[k+*1*]−z*2*[k]z*1*[k+*1*]=s*(*z*1*[k+*1]^{2} *+z*2*[k+*1]^{2})+(−*L*4·*z*2*[k+*1*]+L*5·*z*1*[k+*1])*e[k]* (23)

In the structure of a current observer, sin(ωT) should be estimated from the equation (23). More specifically, the disturbance frequency or the disturbance angular frequency ω that is estimated in an analog operation is not estimated, but sin(ωT) should be directly estimated in a digital control operation. According to the equation (23), sin(ωT) is obtained from the following equation (24):

In the current observer, cos(ωT) may be estimated, instead of sin(ωT). More specifically, the equation (25) is obtained by multiplying z1[k] of the equation (21) by z1[k+1], and multiplying z2[k] of the equation (21) by z2[k+1].

The equation (26) is obtained by summing up the upper and lower sides of the equation (25).

*z*1*[k]z*1*[k+*1*]+z*2*[k]z*2*[k+*1*]=c*(*z*1*[k+*1]^{2} *+z*2*[k+*1]^{2})−(−*L*4·*z*1*[k+*1*]−L*5·*z*2*[k+*1])*e[k]* (26)

The disturbance frequency or the disturbance angular frequency ω that is estimated in an analog operation is not estimated, but cos(ωT) should be directly estimated in a digital control operation. According to the equation (26), cos(ωT) is obtained from the following equation (27).

In the first embodiment, on the other hand, the oscillation frequency is estimated through sin(ωT). Accordingly, in the digital control, the oscillation frequency can be accurately estimated, and position control adapting to the oscillation frequency component can be performed.

Where sin(ωT) is used as the estimated value of the oscillation frequency f in the digital control, the frequency expressed by the following equation (28) is estimated, with Fs representing the sampling frequency in the digital control.

As illustrated in

Where cos(ωT) is used, the curve becomes almost flat in the low frequency zone, and the frequency estimation accuracy becomes poorer. For example, in a case of a sample frequency of 12 kHz, the curve is flat in the low frequency zone around 500 Hz, and the estimation accuracy becomes lower.

Accordingly, sin(ωT) should be used as the estimated value where priority is put on the estimation accuracy on the lower frequency side, and cos(ωT) should be used as the estimated value where priority is put on the estimation accuracy on the higher frequency side.

Furthermore, when the disk apparatus is in use, external oscillations often become a problem on the lower frequency side. For example, external oscillations of several hundreds Hz are often caused from outside. Therefore, in the disk apparatus, it is more convenient to use the estimated value sin(ωT), with which lower frequency oscillations can be estimated with higher accuracy.

The estimator **24** of

In a first adaptive control operation, e[k] of the equation (24) or the equation (27) becomes “0”, if accurate estimation is performed, and the control system can suppress the disturbance frequency. Accordingly, the second term of the right side of each of the equations (24) and (27) should become “0”.

Therefore, where sin(ωT) is estimated, the second term E[k] of the right side of the equation (24) is calculated according to the following equation (29).

Next, E[k] of the equation (29) is multiplied by the adaptive gain K, so that sin(ωT) is updated as in the following equation (30).

sin(ω[*k+*1*]T*)=sin(ω[*k]T*)+*K·E[k]* (30)

Where cos(ωT) is estimated, the second term E[k] of the right side of the equation (27) is calculated according to the following equation (31).

Next, E[k] of the equation (31) is multiplied by the adaptive gain K, so that cos(ωT) is updated as in the following equation (32).

cos(ω[*k+*1*]T*)=cos(ω[*k]T*)+*K·E[k]* (32)

In the above described manner, sin(ωT) and cos(ωT) can be adaptively estimated with the use of the five values, z1[k+1], z2[k+1], e[k], L4, and L5.

A calculation in a second adaptive control operation is now described. Where a fixed-point processor is used, the number of digits is limited. If the above five values, z1[k+1], z2[k+1], e[k], L4, and L5, are used, the number of digits in each of the values of z1[k+1] and z2[k+1] differs from the number of digits in each of the values of e[k], L4, and L5. To effectively use the limited number of digits, calculations are performed in the following manner.

Where sin(ωT) is estimated, the first term G[k] of the right side of the equation (24) is calculated according to the following equation (33).

The difference E[k] between the current set value sin(ωT) and the equation (33) is calculated according to the following equation (34).

*E[k]=G[k]−*sin(ω[*k]T*) (34)

E[k] of the equation (34) is multiplied by the adaptive gain K, so that sin(ωT) is updated as in the following equation (35).

sin(ω[*k+*1*]T*)=sin(ω[*k]T*)+*K·E[k]* (35)

Where cos(ωT) is estimated, the first term G[k] of the right side of the equation (27) is calculated according to the following equation (36):

The difference E[k] between the current set value cos(ωT) and the equation (36) is calculated according to the following equation (37).

*E[k]=G[k]−*cos(ω[*k]T*) (37)

E[k] of the equation (37) is multiplied by the adaptive gain K, so that cos(ωT) is updated as in the following equation (38).

cos(ω[*k+*1*]T*)=cos(ω[*k]T*)+*K·E[k]* (38)

In the above described manner, sin(ωT) and cos(ωT) can be adaptively estimated only with the use of the four values, z1[k+1], z1[k], z2[k+1], and z2[k].

Next, a method for eliminating the errors due to the adaptive gain K in the equations (30), (32), (35), and (38) is described. Where the disturbance frequency or the disturbance angular frequency is determined through analog control, the adaptive law is represented by the equation (6). The error value in the equation (6) contains a noise component other than a disturbance (such as Repeatable Run Out (RRP) and Non Repeatable Run Out (NRRO).

A case where the noise influence exhibits the same value with respect to a disturbance at any frequency is described. In a disturbance frequency following state, the variation of the estimated frequency is determined by the variations of the adaptive gain K and the error value, and accordingly, is within a certain range at any disturbance frequency, as long as the same adaptive gain K is used.

In the digital control, on the other hand, the error value E[k] is determined, and is multiplied by the adaptive gain K, so as to update sin(ωT), as in the equations (30), (32), (35), and (38). In those equations, the adaptive gain is a fixed value.

In the following, a case where the error value E[k] varies due to noise, and the value of the variation is the same at any disturbance frequency is described. The influence of the variation of the error value E[k] has on ω[k] when the disturbance frequency is low or sin(ωT) is close to “0” differs from the influence of the variation of the error value E[k] has on ω[k] when the disturbance frequency is high or sin ωT is close to “1”.

As illustrated in

Where the variation width varies as above, the influence on errors becomes large, if the adaptive gain in digital control is the same. Therefore, the adaptive gain should vary in accordance with the current estimated disturbance frequency. More specifically, the variable adaptive gain should be large at a low frequency, and should be small at a high frequency.

Accordingly, in a process of estimating sin(ωT), the adaptive gain Ka is a value that is proportional to cos(ωT), as indicated by the following equation (40). In this manner, the influence of the difference in variation width on errors can be restrained.

*Ka=K·*cos(ω*T*) (40)

Likewise, in a process of estimating cos(ωT), the adaptive gain Ka is a value that is proportional to sin(ωT), as in the equation (40).

The compensator **23** illustrated in

The control variable table **25** of

Accordingly, for each sample, the values of L4, L5, and cos(ωT) or sin(ωT) corresponding to sin(ωT) or cos(ωT) calculated by the estimator **24** are read from the control variable table **25**, and are set into the compensator **23**. Based on the set control constants and the position error, the compensator **23** calculates the disturbance suppression control amount Udist according to the equation (41), and outputs the disturbance suppression control amount Udist to the addition module **26**.

**21** and the compensator **23** are formed with an integral current observer. Also, in this example, the sin(ωT) estimator **24** utilizes the calculation method of the first adaptive control.

The position control system of

In the equation (44), a11 through a22 represent the values of sine and cosine in the equations (9) and (18).

In other words, the second embodiment is an example of a position control system that is formed with an observer involving a controller model and a disturbance model. In

In **20** subtracts the target position r from an observed position y[k] obtained by demodulating the above described servo information read by the head **3** (see

A second arithmetic module **32** subtracts an estimated position x[k] of the observer from the actual position error er[k] to determine an estimated position error e[k]. A state estimating module **34** uses the estimated position error e[k] and the estimated gains L (L1, L2, L3, L4, and L5) as interval variables of the equation (42), to determine estimated correction values (the second term of the right side of the equation (42)).

An addition module **36** adds the estimated correction value to the state quantities from a delay module **46** (the first terms of the right side of the equation (42)) x[k], v[k], b[k], z1[k], and z2[k], so as to determine an estimated position x[k], an estimated velocity v[k], an estimated bias value b[k], and estimated disturbance suppression values z1[k] and z2[k] as in the equation (42). In the equation (44), the estimated position error e[k] is represented by (y[k]−x[k]).

The estimated values are multiplied by a state feedback gain (−Fa=F1, F2, F3, F4, and F5) at a fourth arithmetic module **38**, to obtain the driving value u[k] of the actuator **1**. Meanwhile, the estimated value of the equation (42) from the addition module **36** is multiplied by an estimated gain Aa (the matrix of the first term of the right side of the equation (44)) at a fifth arithmetic module **42**.

The driving value u[k] of the fourth arithmetic module **38** is multiplied by an estimated gain B (by which u[k] of the second term of the right side of the equation (44) is multiplied) at a sixth arithmetic module **40**. The two multiplication results are added to each other at an addition module **44**, to obtain estimated state quantities x[k+1], v[k+1], b[k+1], z1[k+1], and z2[k+1] of the next sample of the equation (44).

As described above, the estimated state quantities of the next samples are input to the delay module **46**, and are corrected with the estimated correction values at the state estimating module **34**. At a seventh arithmetic module **48**, the estimated position x[k] is extracted from the estimated value of the equation (42) from the addition module **36**, and is input to the second arithmetic module **32**.

The estimator **24** and the control variable table **25** are given to this current observer. The estimator **24** calculates sin(ωT) according to the equations (29) and (30), which are the first adaptive law of sin(ωT).

More specifically, in the estimator **24**, an estimated error (E[k]) calculator **24**-**1** receives the estimated position error e[k], the estimated state quantities z1[k+1] and z2[k+1], and the estimated gains L4 and L5. According to the equation (29), the estimated error calculator **24**-**1** calculates and outputs an estimated error E[k].

A gain multiplier **24**-**2** multiplies the estimated error E[k] by a gain Ka, and calculates the second term of the right side of the equation (30). An adder **24**-**5** adds sin(ω[k]T) of the previous sample from a delay module **24**-**3** to the output of the gain multiplier **24**-**2**, and outputs sin(ω[k+1]T) of the equation (30). The sin(ω[k+1]T) is delayed by one sample at the delay module **24**-**3**, and is input to the adder **24**-**5**.

If the input sin(ωT) is smaller than “0”, a clip module (a saturation processing module) **24**-**4** restricts sin(ωT) to “0”. If the input sin(ωT) is larger than “1”, the clip module **24**-**4** restricts sin(ωT) to “1”. Accordingly, even if sin(ωT) is estimated from the state variables, a disturbance frequency component within the frequency range illustrated in

As described above, the values of sin(ωT) corresponding to the frequency range of 0 to Fs/4 fall into the range of 0 to 1. Therefore, the estimated value of sin(ωT) needs to be always restricted within the range of 0 to 1. In the second embodiment, the clip module **24**-**4** is provided to restrict the output of the addition module **24**-**5** within the range of 0 to 1.

A gain varying module **24**-**6** calculates cos(ω[k]T) from sin(ω[k]T) supplied from the delay module **24**-**3**, and updates the gain Ka of the gain multiplier **24**-**2** according to the equation (40).

The output sin(ω[k+1]T) of the estimator **24** is input to the control variable table **25**. As illustrated in **25** stores the values of the estimated gains L1, L2, L3, L4, and L5, and the values of cos(ωT) corresponding to the estimated values of sin(ωT). As illustrated in

The control variable table **25** extracts the estimated gain L and cos(ωT) corresponding to the estimated value of sin(ωT) and sets the extracted values in the modules **34** and **42**. In this manner, the estimated gains L1 through L5 of the equation (42) of the module **34** and the control constants a11 through a22 (=cos(ωT) or sin(ωT)) of the equation (44) of the module **42** are updated, and the disturbance suppression characteristics of the observer are changed in accordance with the disturbance frequency.

As described above, where a current observer is used, designing is easy, and the disturbance suppression function can be readily set. Also, in a case where the estimator **24** estimates cos(ωT), the same structure as above will do. In a case where cos(ωT) is estimated, the clip module **24**-**4** restricts the clip range between “−1” and “+1”, so as to achieve the same effects.

**24**A in the structure illustrated in

In **24**A calculates sin(ωT) according to the equations (33) through (35), which are the second adaptive law of sin(ωT). More specifically, in the estimator **24**A, an estimated error (G[k]) calculator **24**-**7** receives the estimated state quantities z1[k], z1[k+1], z2[k], and z2[k+1]. According to the equation (33), the estimated error calculator **24**-**7** calculates and outputs a first estimated error G[k].

A first arithmetic module **24**-**8** subtracts sin(ω[k]T) of the previous sample supplied from the delay module **24**-**3** from the first estimated error G[k], and outputs a second estimated error E[k] of the equation (34).

The gain multiplier **24**-**2** multiplies the second estimated error E[k] by a gain Ka, and calculates the second term of the right side of the equation (35). The adder **24**-**5** adds sin(ω[k]T) of the previous sample from the delay module **24**-**3** to the output of the gain multiplier **24**-**2**, and outputs sin(ω[k+1]T) of the equation (35). The sin(ω[k+1]T) is delayed by one sample at the delay module **24**-**3**, and is input to the first arithmetic module **24**-**8** and the adder **24**-**5**.

If the input sin(ωT) is smaller than “0”, the clip module **24**-**4** restricts sin(ωT) to “0”. If the input sin(ωT) is larger than “1”, the clip module **24**-**4** restricts sin(ωT) to “1”.

The gain varying module **24**-**6** then calculates cos(ω[k]T) from sin(ω[k]T) supplied from the delay module **24**-**3**, and updates the gain Ka of the gain multiplier **24**-**2** according to the equation (40).

The output sin(ω[k+1]T) of the estimator **24**A is input to the control variable table **25**, as in the example illustrated in **25** extracts the estimated gain L and cos(ωT) corresponding to the estimated value sin(ωT), and sets the extracted values in the modules **34** and **42**. In this manner, the estimated gains L1 through L5 of the equation (40) of the module **34** and the control constants a11 through a22 (=cos(ωT) or sin(ωT)) of the equation (42) of the module **42** are updated, and the disturbance suppression characteristics of the observer are changed in accordance with the disturbance frequency.

As described above, where a current observer is used, designing is easy, and the disturbance suppression function can be readily set. Since the second adaptive law is used, there is not a difference in the number of digits among inputs. Even if a fixed-point processor is used, calculations can be performed with high precision. Furthermore, in a case where the estimator **24**A estimates cos(ωT), the equations (36) through (38) can be calculated with the same structure.

In the above described frequency estimation, sin(ωT) is estimated. Meanwhile, depending on the purpose of use, there is a demand for a wider estimated frequency range. For example, there is a demand for an expansion to Fs/2, which is a Nyquist frequency. The expansion of the frequency range is described below.

In **27** estimates sin(ωT/2) as a disturbance frequency. A control variable table **28** stores the control constants corresponding to the values of sin(ωT/2), as in the example illustrated in

The values z1[k+1] and z2[k+1] according to the adaptive law (**24**) of the current observer are merely estimated values. Therefore, estimated values z1[k+0.5] and z2[k+0.5] are created by delaying the estimated values z1[k+1] and z2[k+1] by half a sample.

Since z1[k+1] and z2[k+1] are the sine and cosine components of the rotation vectors of the oscillation frequency, the rotation vector should be delayed by ωT/2, to delay z1[k+1] and z2[k+1] of the k+1 sample by half a sample, or create the rotation vectors z1[k+0.5] and z2[k+0.5] of the k+0.5 sample. Accordingly, z1[k+0.5] and z2[k+0.5] delayed by half a sample can be calculated from z1[k+1] and z2[k+1] according to the following equation (45).

The left side and the right side of this equation (45) are replaced with each other, and a conversion is performed to express z1[k+1] and z2[k+1] independently of each other. As a result, the following equation (46) is obtained.

The equation (47) is obtained by substituting the equation (46) into the equation (17) of z1[k+1] and z2[k+1].

To further simplify the equation, the left side and the right side of the equation (47) is multiplied by the following expression (48).

At this point, the following equations (49) are obtained through the multiplications between the cosine and sine matrixes.

With the use of the equations (49), the left side and the right side of the equation (47) are multiplied by the equation (48), to obtain the following equation (50).

In other words, a relational expression of sin(ωT/2) is obtained, as in the equation (17) of sin(ωT). The equation (50) is expanded as in the estimation of sin(ωT). More specifically, by simplifying sin(ωT/2) and cos(ωT/2) to “s” and “c” the equation (50) is transformed into the equation (51).

The equation (52) is obtained by determining z1[k] and z2[k] of the k sample from the equation (51).

The equation (53) is obtained by multiplying the upper side of the equation (52) by z2[k+0.5], and multiplying the lower side of the equation (52) by −z1[k+0.5]. Here, the left-side negative sign “−” is given to z2[k].

The equation (54) is obtained by summing up the upper and lower sides of the equation (53).

*z*1*[k]z*2*[k+*0.5*]−z*2*[k]z*1*[k+*0.5*]=s*(*z*1*[k+*0.5]^{2} *+z*2*[k+*0.5]^{2})+(−*L*4·*z*2*[k+*0.5*]+L*5·*z*1*[k+*0.5])*e[k]* (54)

According to the equation (54), sin(ωT/2) is expressed by the following equation (55). This equation is used in the adaptive law.

According to the equation (55), sin(ωT/2) is estimated, since the disturbance frequency is directly estimated in a digital control operation.

As illustrated in

Accordingly, the estimation accuracy on the lower frequency side is made higher, while higher frequencies can be estimated. Also, when the disk apparatus is in use, external oscillations often become a problem on the lower frequency side. For example, external oscillations of several hundreds Hz are often caused from outside. Therefore, in the disk apparatus, it is more convenient to use the estimated value sin(ωT/2), with which lower frequency oscillations can be estimated with higher accuracy.

The sin(ωT/2) estimator **27** of

In a first adaptive control operation, e[k] of the equation (55) becomes “0”, if accurate estimation is performed, and the control system can suppress the disturbance frequency. Accordingly, the second term of the right side of the equation (55) should become “0”.

Therefore, where sin(ωT/2) is estimated, the second term E[k] of the right side of the equation (55) is calculated according to the following equation (57).

Next, E[k] of the equation (57) is multiplied by the adaptive gain K, so that sin(ωT/2) is updated as in the following equation (58).

sin(ω[*k+*1*]T/*2)=sin(ω[*k]T/*2)+*K·E[k]* (58)

In the above described manner, sin(ωT/2) can be adaptively estimated with the use of the five values, z1[k+1], z2[k+1], e[k], L4, and L5.

A calculation method in a second adaptive control operation is now described. Where a fixed-point processor is used, the number of digits is limited. If the above five values, z1[k+1] z2[k+1], e[k], L4, and L5, are used, the number of digits in each of the values of z1[k+1] and z2[k+1] differs from the number of digits in each of the values of e[k], L4, and L5. To effectively use the limited numbers of digits, calculations are performed in the following manner.

Where sin(ωT) is estimated, the first term G[k] of the right side of the equation (55) is calculated according to the following equation (59).

The difference E[k] between the current set value sin(ωT/2) and the equation (59) is calculated according to the following equation (60).

*E[k]=G[k]−*sin(ω[*k]T/*2) (60)

E[k] of the equation (60) is multiplied by the adaptive gain K, so that sin(ωT/2) is updated as in the following equation (61).

sin(ω[*k+*1]*T/*2)=sin(ω[*k]T/*2)+*K·E[k]* (61)

In the above described manner, sin(ωT/2) can be adaptively estimated only with the use of the four values, z1[k+1] z1[k], z2[k+1], and z2[k].

As in the first embodiment, the adaptive gain is a fixed value in those equations. As in the case of sin(ωT) estimation, the variation of sin(ωT/2) with respect to the variation of ω is small when ω is large, and the variation of sin(ωT/2) with respect to the variation of ω is large when ω is small, as illustrated in

Accordingly, in a process of estimating sin(ωT/2), the adaptive gain Ka is a value that is proportional to cos(ωT/2) as indicated by the following equation (62). In this manner, the influence of the difference in variation width on errors can be restrained.

*Ka=K·*cos(ω*T/*2) (62)

The compensator **23** illustrated in

The control variable table **28** of

Accordingly, for each sample, the values of L4, L5, cos(ωT), and sin(ωT) corresponding to sin(ωT/2) calculated by the sin(ωT/2) estimator **27** are read from the control variable table **28**, and are set into the compensator **23**. Based on the set control constants and the position error, the compensator **23** calculates the disturbance suppression control amount Udist according to the equation (41), and outputs the disturbance suppression control amount Udist to the addition module **26**.

As described above, where position control is performed in accordance with the disturbance frequency in the digital position control system, the disturbance frequency is estimated with high precision, and position control adapting to the disturbance frequency can be performed.

**21** and the compensator **23** are formed with an integral current observer. Also, in the example, the estimator **27** utilizes the calculation method of the first adaptive control.

The position control system of

In **21**, **23**, and **26** are the same as those illustrated in **27** and the control variable table **28** are given to the current observer. The estimator **27** calculates sin(ωT/2) according to the equations (57) and (58), which is the first adaptive law of sin(ωT/2).

More specifically, in the estimator **27**, a state point converting module **27**-**1** converts z1[k+1] and z2[k+1] into z1[k+0.5] and z2[k+0.5], according to the equation (45). An estimated error calculator **27**-**2** receives the estimated position error e[k], the estimated state quantities z1[k+0.5] and z2[k+0.5], and the estimated gains L4 and L5. According to the equation (57), the estimated error calculator **27**-**2** calculates and outputs an estimated error E[k].

A gain multiplier **27**-**3** multiplies the estimated error E[k] by the gain K, and calculates the second term of the right side of the equation (58). An adder **27**-**5** adds sin(ω[k]T/2) of the previous sample from a delay module **27**-**4** to the output of the gain multiplier **27**-**3**, and outputs sin(ω[k+1]T/2) of the equation (58). The sin(ω[k+1]T/2) is delayed by one sample at the delay module **27**-**4**, and is input to the adder **27**-**5**.

If the input sin(ωT/2) is smaller than “0”, a clip module (a saturation processing module) **27**-**6** restricts sin(ωT/2) to “0”. If the input sin(ωT) is larger than “1”, the clip module **24**-**4** restricts sin(ωT/2) to “1”.

A gain varying module **27**-**7** calculates cos(ω[k]T/2) from sin(ω[k]T/2) supplied from the delay module **27**-**4**, and updates the gain Ka of the gain multiplier **27**-**3** according to the equation (62).

The output sin(ω[k+1]T/2) of the estimator **27** is input to the control variable table **28**. As illustrated in **28** stores the values of the estimated gains L1, L2, L3, L4, and L5, and the values of cos(ωT) and sin(ωT) corresponding to the estimated values of sin(ωT/2). As illustrated in

The control variable table **28** extracts the estimated gain L, cos(ωT), and sin(ωT) corresponding to the estimated value sin(ωT/2), and sets the extracted values in the modules **34** and **42**. In this manner, the estimated gains L1 through L5 of the equation (42) of the module **34** and the control constants a11 through a22 (=cos(ωT) or sin(ωT)) of the equation (44) of the module **42** are updated, and the disturbance suppression characteristics of the observer are changed in accordance with the disturbance frequency.

As described above, where a current observer is used, designing is easy, and the disturbance suppression function can be readily set.

**27** of **27**A utilizes the calculation method of the second adaptive control.

In **27**A calculates sin(ωT/2) according to the equations (59) through (61), which are the second adaptive law of sin(ωT/2).

More specifically, in the estimator **27**A, the state point converting module **27**-**1** converts z1[k+1] and z2[k+1] into z1[k+0.5] and z2[k+0.5], according to the equation (45). An estimated error (G[k]) calculator **27**-**8** receives the estimated state quantities z1[k], z1[k+0.5], z2[k], and z2[k+0.5]. According to the equation (59), the estimated error calculator **27**-**8** calculates and outputs a first estimated error G[k].

A first arithmetic module **27**-**9** subtracts sin(ω[k]T/2) of the previous sample supplied from the delay module **27**-**4** from the first estimated error G[k], and outputs a second estimated error E[k] of the equation (60).

The gain multiplier **27**-**3** multiplies the second estimated error E[k] by the gain K, and calculates the second term of the right side of the equation (61). The adder **27**-**5** adds sin(ω[k]T/2) of the previous sample from the delay module **27**-**4** to the output of the gain multiplier **27**-**3**, and outputs sin(ω[k+1]T/2) of the equation (61). The sin(ω[k+1]T/2) is delayed by one sample at the delay module **27**-**4**, and is input to the arithmetic module **27**-**9** and the adder **27**-**5**.

If the input sin(ωT/2) is smaller than “0”, the clip module (a saturation processing module) **27**-**6** restricts sin(ωT/2) to “0”, as in the example illustrated in **27**-**6** restricts sin(ωT/2) to “1”, as in the example illustrated in

The gain varying module **27**-**7** then calculates cos(ω[k]T/2) from sin(ω[k]T/2) supplied from the delay module **27**-**4**, and updates the gain Ka of the gain multiplier **27**-**3** according to the equation (62).

The output sin(ω[k+1]T/2) of the estimator **27**A is input to the control variable table **28**. The control variable table **28** is the same as that illustrated in **28** extracts the estimated gain L, sin(ωT), and cos(ωT) corresponding to the estimated value sin(ωT/2), and sets the extracted values in the modules **34** and **42**. In this manner, the estimated gains L1 through L5 of the equation (42) of the module **34** and the control constants a11 through a22 (=cos(ωT) or sin(ωT) of the equation (44) of the module **42** are updated, and the disturbance suppression characteristics of the observer are changed in accordance with the disturbance frequency.

As described above, where a current observer is used, designing is easy, and the disturbance suppression function can be readily set. Since the second adaptive law is used, there is not a difference in the number of digits among inputs. Even if a fixed-point processor is used, calculations can be performed with high precision.

**21** and the compensator **23** are formed with separate current observers. Also, in this example, the estimator **24** utilizes the calculation method of the first adaptive control.

The position control system of

In the equation (67), a11 through a22 represent the values of sine and cosine expressed by the equations (9) and (18).

In other words, the fifth embodiment is an example of a position control system that is formed with observers including a controller model and a disturbance model. In **9**, and **16** are denoted by the same reference numerals and symbols as those used in **9**, and **16**.

In **20** subtracts a target position r from an observed position y[k] obtained by demodulating the above described servo information read by the head **3** (see

The second arithmetic module **32** subtracts an estimated position x[k] of the observer from the actual position error er[k], to determine an estimated position error e[k]. A state estimating module **34**-**1** uses the estimated position error e[k] and the estimated gains L (L1 and L2) as interval variables of the equation (63), to determine estimated correction values (the second term of the right side of the equation (63)).

An addition module **36**-**1** adds the estimated correction value to the state quantities from a delay module **46**-**1** (the first terms of the right side of the equation (63)) x[k] and v[k], so as to determine an estimated position x[k] and an estimated velocity v[k]. In the equation (63), the estimated position error e[k] is represented by (y[k]−x[k]).

The estimated values are multiplied by a state feedback gain (−Fa=F1 and F2) at a third arithmetic module **38**-**1**, to obtain the driving value u[k] of the actuator **1**. Meanwhile, the estimated value of the equation (63) from the addition module **36**-**1** is multiplied by an estimated gain Aa (the matrix of the first term of the right side of the equation (66)) at a fourth arithmetic module **42**-**1**.

The driving value u[k] of the third arithmetic module **38**-**1** is multiplied by an estimated gain B (by which u[k] of the second term of the right side of the equation (66) is multiplied) at a fifth arithmetic module **40**. The two multiplication results are added to each other at the addition module **44**, to obtain estimated state quantities x[k+1] and v[k+1] of the next sample of the equation (66).

As described above, the estimated state quantities of the next samples are input to the delay module **46**-**1**, and are corrected with the estimated correction values at the state estimating module **34**-**1**. At a sixth arithmetic module **48**, the estimated position x[k] is extracted from the estimated value of the equation (63) from the addition module **36**-**1**, and is input to the second arithmetic module **32**.

Likewise, in the disturbance model, a second state estimating module **34**-**2** uses the estimated position error e[k] and the estimated gains Ld1 (L3, L4, and L5) as interval variables of the equation (63), to determine estimated correction values (the second term of the right side of the equation (63)).

A second addition module **36**-**2** adds the estimated correction value to the state quantities from a second delay module **46**-**2** (the first terms of the right side of the equation (63)) b [k], z1[k], and z2[k], so as to determine an estimated bias b[k] and estimated disturbance quantities z1[k] and z2[k].

The estimated values are multiplied by a state feedback gain (−Fd1=F3, F4, and F5) at a seventh arithmetic module **38**-**2**, to obtain the disturbance suppression driving value ud[k] of the actuator **1**. Meanwhile, the estimated value of the equation (63) from the second addition module **36**-**2** is multiplied by an estimated gain Ad1 (the matrix of the right side of the equation (67)) at an eighth arithmetic module **42**-**2**.

A ninth arithmetic module **50** adds the driving value ud[k] of the seventh arithmetic module **38**-**2** to the driving value u[k] of the third arithmetic module **38**-**1**, and outputs the driving value u[k] of the actuator.

In this example, the estimator **24** and the control variable table **25** are also given to the current observers. The estimator **24** is the same as that illustrated in **25** is the same as that illustrated in

The control variable table **25** sets the estimated gains L1 and L2 corresponding to the estimated value sin(ωT) in the state estimating module **34**-**1**, and sets the estimated gains L3, L4, and L5, and sin(ωT) and cos(ωT) in the modules **34**-**2** and **42**-**2**. In this manner, the estimated gains L1 through L5 of the equation (63) of the modules **34**-**1** and **34**-**2**, and the control constants a11 through a22 (=cos(ωT) or sin(ωT)) of the equation (67) of the module **42**-**2** are updated, and the disturbance suppression characteristics of the observers are changed in accordance with the disturbance frequency.

As described above, where a disturbance model is formed with an independent current observer, the designing of the disturbance model is easy, and the disturbance suppression function can be readily set.

**21** and the compensator **23** are formed with separate current observers. Also, in the sixth embodiment, the sin(ωT) estimator **24**A utilizes the calculation method of the second adaptive control.

In

In other words, the sixth embodiment is an example of a position control system that is formed with observers including a controller model and a disturbance model, like the example illustrated in

In the sixth embodiment, the estimator **24**A and the control variable table **25** are also given to the current observers. The estimator **24**A is the same as that illustrated in **25** is the same as that illustrated in

The control variable table **25** sets the estimated gains L1 and L2 corresponding to the estimated value sin(ωT/2) in the state estimating module **34**-**1**, and sets the estimated gains L3, L4, and L5, and sin(ωT) and cos(ωT) in the modules **34**-**2** and **42**-**2**. In this manner, the estimated gains L1 through L5 of the equation (63) of the modules **34**-**1** and **34**-**2**, and the control constants a11 through a22 (=cos ωT or sin ωT) of the equation (67) of the module **42**-**2** are updated, and the disturbance suppression characteristics of the observers are changed in accordance with the disturbance frequency.

As described above, where a disturbance model is formed with an independent current observer, the designing of the disturbance model is easy, and the disturbance suppression function can be readily set.

In the description with reference to

Also, in the above descriptions, the compensator **23** is formed with an observer. However, the compensator **23** of

Although the above position control apparatuses are designed for heads, the present invention may also be applied to other position control apparatuses. Likewise, although the disk apparatuses are magnetic disk apparatuses in the above embodiments, the present invention may also be applied to other data storage devices.

As mentioned above, a value of one of estimated errors of sin(ωT), cos(ωT), and sin(ωT/2) is calculated with ω representing an angular frequency of a disturbance and T representing the sampling period. Then, the calculated estimated error is multiplied by an adaptive gain, and one of the values sin(ωT), cos(ωT), and sin(ωT/2) is updated, so as to perform the disturbance adaptive control. Therefore, the estimation accuracy of the disturbance frequency can be improved, thereby the disturbance adaptive control can be realized with high accuracy.

The various modules of the systems described herein can be implemented as software applications, hardware and/or software modules, or components on one or more computers, such as servers. While the various modules are illustrated separately, they may share some or all of the same underlying logic or code.

While certain embodiments of the inventions have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel methods and systems described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.

Patent Citations

Cited Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|

US5708581 | Jul 12, 1996 | Jan 13, 1998 | Hewlett-Packard Company | Method for maximizing feedforward orthogonality for minimizing servo system nuller instability |

US6421200 | Dec 27, 2000 | Jul 16, 2002 | Fujitsu Limited | Control method and control device for a disk storage device |

US6636376 * | Oct 12, 2000 | Oct 21, 2003 | Seagate Technology Llc | Disc drive resonance disturbance cancellation filter |

US7265934 | Feb 23, 2006 | Sep 4, 2007 | Fujitsu Limited | Head position control method, head position control device, and disk device |

US7372659 * | Nov 21, 2006 | May 13, 2008 | Fujitsu Limited | Head position control method, head position control device and disk device |

US20060114602 | Nov 8, 2005 | Jun 1, 2006 | Samsung Electronics Co., Ltd. | Non-repeatable run out compensating apparatus and method and disc drive using the same |

US20060291101 | Feb 23, 2006 | Dec 28, 2006 | Fujitsu Limited | Head position control method, head position control device, and disk device |

US20080065240 * | Apr 25, 2007 | Mar 13, 2008 | Fujitsu Limited | Position control method, position control device, and medium storage device having disturbance suppression function |

US20100014185 * | Jan 21, 2010 | Fujitsu Limited | Position control apparatus and disk apparatus using the same | |

US20100014186 * | Sep 24, 2009 | Jan 21, 2010 | Fujitsu Limited | Position control apparatus and disk apparatus using the same |

US20100268354 * | Oct 21, 2010 | Toshiba Storage Device Corporation | Position control method, position control device, and medium storage device having disturbance suppression function | |

JP2526305B2 | Title not available | |||

JP3460795B2 | Title not available | |||

JP3819114B2 | Title not available | |||

JP2005025880A | Title not available | |||

JP2006155875A | Title not available | |||

JP2007004946A | Title not available | |||

JPH0750075A | Title not available |

Referenced by

Citing Patent | Filing date | Publication date | Applicant | Title |
---|---|---|---|---|

US8804274 * | Sep 4, 2013 | Aug 12, 2014 | Kabushiki Kaisha Toshiba | Disk storage apparatus and servo control method |

Classifications

U.S. Classification | 360/77.02, 360/78.04, 360/77.04 |

International Classification | G11B5/596 |

Cooperative Classification | G11B5/5582, G11B5/59627, G11B5/5547 |

European Classification | G11B5/55D6, G11B5/55D1D2 |

Legal Events

Date | Code | Event | Description |
---|---|---|---|

Sep 23, 2009 | AS | Assignment | Owner name: FUJITSU LIMITED,JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:TAKAISHI, KAZUHIKO;REEL/FRAME:023274/0550 Effective date: 20090827 Owner name: FUJITSU LIMITED, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:TAKAISHI, KAZUHIKO;REEL/FRAME:023274/0550 Effective date: 20090827 |

Nov 23, 2009 | AS | Assignment | Owner name: TOSHIBA STORAGE DEVICE CORPORATION,JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:FUJITSU LIMITED;REEL/FRAME:023558/0225 Effective date: 20091014 Owner name: TOSHIBA STORAGE DEVICE CORPORATION, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:FUJITSU LIMITED;REEL/FRAME:023558/0225 Effective date: 20091014 |

Feb 8, 2012 | AS | Assignment | Owner name: KABUSHIKI KAISHA TOSHIBA, JAPAN Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:TOSHIBA STORAGE DEVICE CORPORATION;REEL/FRAME:027674/0653 Effective date: 20120125 |

Feb 11, 2015 | FPAY | Fee payment | Year of fee payment: 4 |

Rotate